Moral decision-making in real life: factors affecting ... - CORE

Moral decision-making in real life:

factors affecting moral orientation and behaviour justification

Shira Haviv

&

Patrick J. Leman

Department of Psychology

Goldsmiths College, University of London

Address for correspondence:

Dr. Patrick Leman

Department of Psychology

Goldsmiths College, University of London

New Cross, London, SE14 6NW

United Kingdom

Tel. +44 (0) 20 7919 7918; Fax. +44 (0) 20 7919 7873

Email. [email protected]

Running Head: Decision-making in real life

Keywords: decision-making, dilemma, gender, moral orientation, moral reasoning

brought to you by COREView metadata, citation and similar papers at core.ac.uk

provided by Royal Holloway Research Online

https://core.ac.uk/display/78866664?utm_source=pdf&utm_medium=banner&utm_campaign=pdf-decoration-v1

ABSTRACT

The study addresses two separate but related issues in connection with peoples’ real life

moral decisions and judgements. First, the notion of moral orientation is examined in terms of

its consistency across varying contexts, its relation to gender and to gender role. Second, a

new aspect of moral reasoning is explored – the influence on moral decision-making of

considering the consequences of an action. Fifty-eight undergraduate students were asked to

discuss two personal and two impersonal real life moral dilemmas. The results reveal a

significant interaction between gender role and type of dilemma. However, moral orientation

was not consistent across various dilemmas and gender was not related to any particular

orientation. Also the results indicate a significant difference between the reasoning of

consequences of personal-antisocial conflicts and impersonal-antisocial conflicts. These

findings suggest that different moral orientations may be embedded in life experience and

connect with an individual’s sense of his or her moral identity in real-life situations.

Moral decision-making in real life / 1

INTRODUCTION

Whilst much work on moral reasoning has sought to identify the relatively abstract

conceptual structures which underpin moral judgements, relatively little has sought to

examine the ways in which moral judgements and decisions are made in real life. There are,

however, good reasons for examining the ways in which these more “everyday” judgements

are made. Not only is it important to learn something more of the social and cultural context

in which moral thought is embedded. It is also crucial to explore moral decision-making in

real life to understand better the links between more hypothetical moral reasoning and moral

action.

This paper presents research that seeks to identify factors that influence the decisions

that individuals make in response to real life moral dilemmas. Its empirical focus will form

two, inter-related strands. An initial strand considers how social contextual factors

(internalised notions of gender role and more externalised notions of gender) relate to moral

orientation. A second strand explores how different types of dilemmas may entail different

consequences for individuals that, in turn, relate to the sorts of justifications and explanations

they give for particular courses of action.

Moral decision-making in real life

A famous example of the difference between hypothetical reasoning and justification of

experienced behaviour is given by Milgram’s (1963) study of obedience. When participants

are faced with a hypothetical dilemma of either harming an innocent stranger or disobeying an

authority figure they frequently choose the latter. However, the study showed that whilst more

than sixty five percent of those who were faced with the dilemma in reality chose to harm an

innocent stranger, few felt such behaviour would be morally acceptable when asked about an

imaginery scenario. Moreover when Milgram’s adult participants were asked about their

actions they justified their behaviour with reasons equivilant to stage one on Kohlberg’s


(1969, 1984) model – for example, the status of the experimenter. It would seem unilkely that

all of Milgram’s particpiants were stage one reasoners. Rather, something about the

experimental situation and their perceived roles within it influenced their moral decisions,

judgements and ultimately their behaviour.

Dominant models of moral development (e.g. Kohlberg, 1969) have focused theoretical

attention on age-related shifts in moral reasoning. Kohlberg’s methodology involved

presenting individuals with various hypothetical moral dilemmas which were either discussed

with the participants by means of an interview or reflected upon using a questionnaire.

Responses are then scored according to a specific manual devised by Kohlberg and his

colleagues. Yet although the value and influence of this work on moral development is clear,

a consistent criticism of, for instance, Kohlberg’s theory has been that he failed adequately to

consider what we might term “real life” moral decision-making (Krebs, Denton & Wark,

1997; Leman, 2001). Research on real life dilemmas (Krebs, Denton, Vermeulen, Carpendale

& Bush, 1991; Carpendale & Krebs 1995; Wark & Krebs 1996, 1997) found that once

participants are asked to judge moral conflicts that they have experienced in their life, moral

stage tends to be lower, and stage consistency of judgements diminishes across different types

of moral dilemmas. A further point of criticism highlighted the difference between moral

judgement competency and moral judgement in practice.

Studies that compare moral behaviour (action) and moral reasoning have highlighted

the problem of how stages of reasoning (derived from hypothetical problems) are related to

real life moral behaviour. Denton and Krebs (1990) found that despite people’s

acknowledgement of the wrongfulness of impaired driving they still drove home while having

high blood alcohol level; Carpendale and Krebs (1995) showed that a monetary incentive also

affected moral choices. And Walker (1984) has claimed that Kohlberg’s stage theory has a

self-limiting scope in that it does not deal directly with the issues of moral emotions and

behaviour – rather, it deals with the adequacy of justifications for solutions to moral conflicts.

Wark and Krebs (1996) summarise a position common to many in arguing that whilst there


are numerous studies on moral judgement only a few have investigated the important and

socially pertinent question of how people make moral decisions in their everyday lives.

Gender and moral orientation

Another critic of Kohlberg’s emphasis on abstract aspects of moral thought was Carol

Gilligan. Gilligan (1982) argued that Kohlberg’s theory is insensitive to the way females view

morality and that there are sex related (but not sex-specific) differences in an individual’s

orientation to life. These differences become particularly visible in terms of moral reasoning.

On one hand, men have a justice orientation which involves an emphasis on autonomy,

separateness and noninterference with abstract rights. On the other women hold a care

orientation involving more emphasis on a concern for the well being of others and a view of

the self as connected and interdependent with others in concrete situations (Walker, de Vries

& Trevethan, 1987).

Gilligan’s evaluation of responses to real life dilemmas by men and women revealed

that although the majority of people used both care and justice orientations, the majority of

women (75%) used a predominantly care orientation whereas the majority of men (79%) used

a predominantly justice orientation. Also, 36% of women did not involve any consideration of

justice in their report and 36% of men did not present any consideration of care. These

findings led Gilligan to conclude that individuals use one predominanting orientation related

to their gender when discussing real life moral conflicts.

Gilligan claimed that males gain higher moral maturity scores on Kohlberg’s test

because they tend to make justice oriented judgments which are captured at higher stages (4-

5). Females, on the other hand, tend to make care oriented judgments which are captured at a

lower stage (3). Although some studies supported this assumption (e.g. Bussey & Maughan,

1982) a large number of studies refuted the claim for significant sex differences in moral

maturity (see Walker, 1984 for a review of the literature). Moreover, research on moral

orientation revealed that studies that found that women use higher percentage of care oriented

terms in real life dilemmas than men were methodologically flawed by not controlling for


type of dilemma (e.g. Ford & Lowery 1986; Walker et al. 1987; Gilligan & Attanucci 1988;

Pratt, Golding, Hunter, & Sampson 1988; Wark & Krebs 1996).

Moral reasoning: considering the consequences of action

Following many studies that compared Kohlberg’s philosophical dilemmas with real

life dilemmas, Krebs and his colleagues have argued that the highest stages of moral

reasoning rarely exist outside the Western academic context (within which Kohlberg’s

philosophical dilemmas were typically tested). In attempting to explore factors that may

explain the variance between judgments (measured by stages in Kohlberg’s model) of

philosophical dilemmas and judgments of real life dilemmas they found an interaction

between moral judgment competency and various performance factors (detailed in Krebs et

al., 1997). Individuals’ ability to retain lower stages of moral judgment and use them in

response to real life dilemmas does not follow Kohlberg’s theoretical assumption regarding

stage replacement, but is more in line with other models of moral judgment such as Rest’s

“layer-cake” model and Levine’s “additive – inclusive” model. Both these models suggest

that new stages are built on old stages, which are retained and may be used in various

circumstances.

Despite this important outcome of recent research, the remaining question is why there

is such descrepency between judgment of philosophical dilemmas and judgment of real life

dilemmas. Krebs et al. (1997) attempted to explain some aspects in real life decision making,

which may serve future research in clarifying how people make moral decisions in their

everyday lives. Two elements are central to the position. One is the distinction between a

third person perspective (which is implied in philosophical dilemmas) and a first person

perspective (which is implied in personal, real life dilemmas). When people come across

moral conflicts in their life the question they are faced with is: “ what should I do?” which is

different to “what should one do?”. Reasoning in real life situations involves decisions, which

are much more practical, self serving, and less rational than reasoning of hypothetical

characters. The second aspect relates to the first in suggesting that factors that people consider


when they make decisions in real life are influenced by functional concerns such as advancing

self-interest or social harmony, and by motivational and affective processes.

One of the most important pragmatic concerns is the consequences of moral decisions.

Krebs et al. (1997) provide a detailed account of the various types of consequences people

consider, which will not be repeated here. However, their explanation of the distinction

between consequences to others and consequences to the self is a central focus of this study.

Although people believe hypothetical characters should act in a certain way and although they

provide reasoning to support that belief, they themselves would have not made that decision

in real life due to the consequences of their decision. For example, despite people’s belief that

Heinz should steal the drug (cf. Kohlberg, 1984, p.640), they themselves would not steal it as

they would not be willing to suffer the consequences (Krebs, Vermeulen, Denton &

Carpendale, 1994). It seems possible that an inconsistency between what one should and

would do in these situations leads to a dissonance that can be partly resolved by changing

one’s mind about a particular course of action (Krebs et al. 1997). Krebs et al. (1997) not only

provide a potential explanation for the inconsistency between reasoning of hypothetical

dilemmas and reasoning of real life moral conflicts. They also point to a direction for future

research into the underlying mechanisms involved in moral judgment, decision, and action.

The present study

The present study had two main aims. First it revisits Gilligan’s distinction between

justice and care moral orientations to examine the links between moral orientation, gender ,

gender role and the type of dilemma under consideration. Despite the ample research that

examined moral orientation, the control for type of real life dilemmas was only recently been

carried out as part of the design rather than a post-hoc analysis. This was done by Wark and

Krebs (1997) who did not find significant sex differences on moral orientation scores. Rather,

they found variations in orientations, which were determined by the type of the dilemmas

(Kohlberg’s dilemmas pulled for justice orintation, real life prosocial dilemmas pulled for

care orientation, & real life antisocial dilemmas pulled for justice orientation). The present


study did not include Kohlberg’s dilemmas but attempted to replicate Wark and Krebs’s

(1997) finding regarding prosocial and antisocial (personal) dilemmas and extend further to

examine impersonal real life dilemmas. The aim of the first part of the study was to explore

whether Gilligan’s model is valid - whether people truly hold a dominant orientation that is

related to their gender. If Gilligan’s model is valid then the predictions are: (a) to find moral

orientation to be consistent across varying contexts, (b) men should make more justice

oriented judgments than women, and (c) women should make more care oriented judgments

than men. However, if Gilligan’s model of moral orientation is not valid then one would

expect to find moral orientation to differ across the various dilemmas (as found by Wark &

Krebs, 1997; Krebs et al. 1994).

The present study also included gender role as a factor. Sochting, Skoe, and Marcia

(1994) found that gender role was a better predictor for the type of orientation one holds. Ford

and Lowery (1986) found that males high on femininity reported higher care ratings than

males low on femininity. Given the large number of studies that found no strict relation

between orientation and gender, a measure for gender role was included to determine whether

gender role will be more predictive of moral orientation than gender.

A second aim of the study involved using an exploratory method to study the aspects

involved in consideration of consequences. Since this second part was exploratory no specific

hypothesis was proposed, but a tentative prediction was that consequences to the self would

differ from consequences to others in their effect on people’s decision-making.

METHOD

Design

The study consisted of two parts. The first related to moral orientation and the second

to one aspect of moral reasoning: consideration of consequences. A mixed design was

employed in the moral orientation part. The three independent variables (IVs) were: gender

(between participants), gender role (between participants), and type of dilemma (within

participants). The dependent variable (DV) was moral orientation score (on a 5-point Likert


scale) which corresponded to the percentage of judgements that were predominantly care-

based.

A within participants design was employed for the second moral reasoning part. The IV

was type of dilemma, and the DV was moral reasoning scores (judged by an independent rater

on a 5-point Likert scale) which represented the degree of influence that participants felt that

considering consequences had on their judgement. For more details on scoring procedures,

see “Scoring” section below.

Participants were allocated to the various experimental conditions based on their gender

and their gender role (as measured by the Personal Attributes Questionnaire–PAQ: Spence &

Helmreich, 1978). Any other allocations were randomly based.

Participants

58 undergraduate students (30 females & 28 males) volunteered for the study. The

average age of males (23 years 7 months) did not differ significantly from the average age of

females (21 years 8 months). There were also no significant differences in age between the

gender role groups.

Socio-economic status (SES) ratings ranging on a scale from one to five were taken on

the basis of parents’ occupation. The mean SES (2.17) did not differ significantly between

men and women. Again, the mean SES for gender role groups were not statistically different.

The average number of years in formal education (14 years) was also not significantly

different between the two gender or gender role groups.

Participants were asked to report whether they have children or not (parental status).

Parental status seems to be an important factor to control following a study by Pratt et al.

(1988) who found that mothers were significantly less justice-oriented than fathers, whereas

the men and women without children were not significantly different. There was no

significant difference between males and females, and between the gender role groups on

parental status (only 4 females and 1 male were parents).


Participants were blind to the aims of the experiment until they returned their

questionnaires.

Materials

Measure of Demographic Information. Participants were given a form to complete the

following details: age, sex, number of years in formal education, work status (whether they

work & name of occupation), parental status, and parents’ occupations.

Gender Role. The Personal Attributes Questionnaire (PAQ: Spence & Helmreich,

1978) was used to distinguish between four gender role types (masculine, feminine,

androgynous, & undifferentiated). The PAQ contains 24 items (adjectives) with a 5-point

Likert scale. All the items refer to socially desirable traits equally divided into three scales

(masculine, feminine, & masculine-feminine). It has been tested for reliability and internal

consistency and rated as highly satisfactory in terms of content, construct, discriminant, and

predictive validity. The PAQ has also been judged as conceptually equivalent to Gilligan’s

self-concept measure (Pratt et al. 1988).

Real Life Moral Dilemmas. Dilemma type has proved to be a very important and

central factor in the study of moral orientation and in research on moral reasoning. Not only

does dilemma content seem to affect the overall score of care based judgments, the

methodology by which dilemmas were chosen and their relation to the participants’ real life

has brought into question the usefulness of the models prevailing in the field (Walker et al.

1987). Although some studies used real-life dilemmas the conflicts in the dilemmas were

raised by the researchers. In more recent research using a self-report method the issues were

raised by the participants themselves. Wark and Krebs (1996) and Krebs et al. (1991)

compared stages of moral judgment between hypothetical dilemmas and real life ones and

found that real life dilemmas led to significantly lower reasoning than that on the hypothetical

dilemmas.

Consequently, the self-report method was employed in the present study. Participants

were asked to describe four real life moral dilemmas followed by probing questions. A


distinction was made between impersonal dilemmas and personal dilemmas. The term

impersonal was defined as “moral conflicts which did not involve you or somebody close to

you directly and that did not necessarily forced you to make a certain decision (to act in a

specific way), but that you consider important and significant.” Whereas the term personal

referred to “moral conflicts which directly involved you and as a result made you take some

sort of action.” Participants were also asked to distinguish between two types of personal

conflicts: antisocial and prosocial. This distinction was taken from Wark and Krebs (1997)1

and was used in order to maintain consistency across dilemmas contents. The present study

used the term antisocial to refer to dilemmas involving “reacting to a transgression (for

example, involving violations of rules, laws or fairness) committed by them; dealing with the

temptation to meet their own needs or desires, acquire resources, or advance their own gain

by violating rules or laws, behaving dishonestly, immorally, or unfairly.” (Wark & Krebs,

1997, p.166). The term prosocial was used to refer to dilemmas involving “dealing with two

or more people making inconsistent demands on them, with implications for their relationship

with each person; deciding whether or not to take responsibility for helping someone

important to them.” (Wark & Krebs, 1997 p.166). These specific dilemma types were selected

by Wark and Krebs (1997) as they were reported most frequently by participants in the Wark

and Krebs (1996) study. Altogether participants were asked to describe two impersonal

dilemmas, one personal antisocial, and one personal prosocial.

Probe Questions. Following the description of each dilemma participants were asked to

answer probing questions. Some of the questions were also based on previous research (Wark

& Krebs, 1997, 2000; Krebs, Denton & Wark, 1997) and included questions such as “what

1 Although Wark and Krebs (1997) used four categories (antisocial: transgression, temptation,

prosocial: helping, loyalty) to define prosocial and antisocial dilemmas, only a broad

distinction was made between the two types in this study (in order to provide participants with

less restricted definitions). Also, it seems a sensible distinction as only one dilemma of each

kind was required, whereas Wark and Krebs (1997) asked for two dilemmas of each type.


did you see to be the issues involved at the time?”, “what made it a moral conflict?”,

“why/why not?”, “what options did you or the person involved in the dilemma consider and

why?”, “do you think you (or the person involved in the dilemma) did the right thing?”

“why/why not?”, “under what circumstances would you have acted differently and why?” The

form contained 22 questions. Participants were asked to describe the dilemma first (in a short

paragraph) and then answer the questions on ensuing pages (appendix 1 contains a list of

probe questions).

Procedure

Participants were given a package of questionnaires containing: (a) a request for

demographic information, (b) the PAQ, (c) a request to describe four moral dilemmas

followed by probing questions. Participants were asked to complete all questionnaires and

return them by a certain date. On returning the questionnaires participants were thanked by

the experimenter and were given an outline of the study.

Scoring

(1) Gender Role. The PAQ was scored as outlined by Spence and Helmreich (1978).

The median scores for the masculine, feminine, and masculine-feminine scales were: 20, 25,

and 18 respectively. Based on these scores participants were categorised as follows: 18

participants (5 females & 13 males) were classified as masculine, 14 participants (9 females

& 5 males) were classified as feminine, 13 participants (8 females & 5 males) were classified

as androgynous, and 13 participants (8 females & 5 males) were classified as undifferentiated.

(2) Moral Orientation. Lyons (1983) provides a scheme for coding considerations of

response (Care) and considerations of rights (Justice). Some considerations of Care were

described as concern with relationships, caring, and the promotion of the welfare of others or

prevention of their harm. Whereas some considerations of Justice were described as

conflicting claims between self and others, and the maintenance of rules, principles, and

standards especially those of fairness and reciprocity. Based on the relative number of care


and justice considerations made to each dilemma, Lyons (1983) classified moral orientation

as either care-based or justice-based. Gilligan and Attanucci (1988) used a refined version of

Lyon’s (1983) scheme in which they classified moral orientation as “care only”, “care focus”,

“care justice”, “justice focus”, and “justice only”. Krebs et al. (1994) used these distinctions

to score judgements on the following 5-point scale: “care only” (C) = 100, “care focus” (C(J))

= 75, “care justice” (C/J) = 50, “justice focus” (J(C)) = 25, and “justice only” (J) = 0.

These distinctions and the 5-point scale were employed in this study. Moral orientation

scores were calculated based on the average of all scored judgements on each dilemma. The

average number of scorable judgements in was 12. The scoring procedure was conducted

blind to all other information about each participant. Thirty percent of each type of dilemma

was scored by a second scorer who was blind to the aim of the study and all other information

about the participants. Inter-rater reliability using the Pearson’s correlation test was 86%

agreement for the impersonal dilemmas (r(16) = 0.857, p< 0.001), 96% agreement for the

prosocial dilemmas (r(16) = 0.927, p< 0.001), and 89% agreement for the antisocial dilemmas

(r(16) = 0.887, p< 0.001). An average score between the two impersonal dilemmas was

calculated for each participant before it was used in the statistical analysis.

Consistency of moral orientation was determined following a procedure outlined in

Walker et al. (1987). Similar to Wark and Krebs (1997) consistency of moral orientation was

assessed in terms of the same or an adjacent score on a 5-point scale (i.e. J, J(C), J/C, C(J), C).

(3) Moral Reasoning – consideration of consequences. A similar but different

procedure to the scoring of moral orientation was used to score the percentage of

consideration of consequences in each dilemma, which resembled the weight these

considerations had on the evaluation of one’s behaviour. Using a 5-point Likert scale ranging

from “consequences did not lead to judgement (of what to do/how to act)” = 0 to

“consequences very much led to judgement (of what to do/how to act)” = 1, all dilemmas

were scored blindly to all other information about each participant. Moral reasoning scores

were calculated based on the scored judgements on each dilemma relative to the overall

number of scorable judgements (the highest number of documented judgements), which was 5


for the impersonal dilemmas and 7 for the personal dilemmas. Thirty per cent of each type of

dilemma was scored by a second scorer who was blind to the aim of the experiment and to all

other information about the participants. The Pearson’s correlation test was used to measure

inter-rater reliability. The test revealed 94% agreement for the impersonal dilemmas, r (16) =

0.937, p<0.001 and 91% agreement for the personal dilemmas, r (16) = 0.908, p<0.001.

RESULTS

The scores of both moral orientation and moral reasoning were recorded on a scale

ranging from 0% to 100%.

Moral Orientation

Means scores for participants’ moral orientation are shown in Table 1. Figure 1 shows

the total means for orientation scores across factors. Higher scores represent higher levels of

care orientation.

--Insert Table 1 about here--

--Insert Figure 1 about here--

As indicated in Table 1, all types of dilemma induced both care-based and justice-based

moral judgements. However, each type pulled with different strength for each orientation. The

prosocial dilemmas had a strong tendency towards the care-based orientation, whereas the

antisocial dilemmas showed strong tendency towards the justice-based orientation (the

impersonal dilemmas appeared to situate around the centre between the two personal

dilemmas). The direction of these results (regarding the prosocial & antisocial dilemmas) is

similar to the findings obtained by Wark and Krebs (1997). However, unlike Wark and Krebs

(1997), the prosocial dilemmas tended towards the care orientation more than the antisocial

tended towards the justice orientation.


A 4 (gender role) X 2 (gender) X 3 (dilemma type) mixed design Anova, with repeated

measures on the last factor was conducted to evaluate the effects of the factors: gender role

(masculine, feminine, androgynous, & undifferentiated), gender (males & females), and

dilemma type (impersonal, antisocial, & prosocial) on moral orientation scores. The test

produced a significant main effect for dilemma type F (2, 100) = 29.96, p<0.001, qualified by

a dilemma type and gender role interaction F (6, 100) = 2.78, p<0.05. There was no

significant main effect for gender and no significant main effect for gender role. All

remaining interactions were also not significant.

These findings do not support Gilligan’s contention since they suggest that the various

contexts of the dilemmas influenced the degree of care responses rather than the dominating

orientation. Furthermore, no clear gender differences were apparent (see again Figure 1).

The data was further analysed in order to find the exact levels at which the interaction

occurred. A series of four within participants Anovas were carried out distinguishing between

the four gender role types, using a Bonferroni adjusted p-value of 0.0125. Results revealed no

significant simple effects for masculine or for undifferentiated groups. However significant

simple effects were found for feminine F (2, 26) = 12.52, p<0.001, and androgynous F (2, 24)

= 19.85, p<0.001.

Three between participants Anovas were carried out to distinguish between the three

types of dilemma – impersonal, antisocial and prosocial. The p-value was adjusted to 0.0167

in line with the appropriate Bonferroni adjustment. Results revealed no significant simple

effects.

Further t-tests were carried out following the simple effects, which were found at both

the feminine and the androgynous conditions of gender role. Impersonal, antisocial, and

prosocial were paired in 3 possible combinations. 3 paired-sample t-tests were conducted at

the feminine gender role condition with an adjusted p-value of 0.0083 (Bonferroni). These

tests revealed a significant difference between the prosocial and the antisocial dilemmas t (13)

= 4.16, p<0.001, and between the prosocial and the impersonal dilemmas t (13) = 4.47,

p<0.001. The pairing of the antisocial and the impersonal dilemmas did not reach


significance. Three further t-tests revealed significant differences between the prosocial and

the antisocial dilemmas at the androgynous level t (12) = 4.63, p<0.001, and between the

prosocial and the impersonal dilemmas t (12) = 9.73, p<0.001. However no significant

difference was found between the antisocial and the impersonal dilemmas.

This outcome supports research on gender role differences and the assumption that

gender role is a better predictor of moral orientation than gender. This study did not find

orientation to differ between males and females. Similar to Ford and Lowery (1986),

femininity level seems to affect the degree of care orientation one holds. This research also

finds a relationship between moral orientation and an androgynous gender role.

Consistency in moral orientation across dilemmas

According to Gilligan (1986), a large proportion of one’s reasoning should reflect

either care or justice orientation with relatively little consideration for the other across various

contexts. The present study failed to support Gilligan’s postulation. Only 3 participants (2

females and 1 male) consistently expressed the same orientation, and none of them scored

exactly the same on all four dilemmas (they were consistent only on adjacent scores).

Moreover, no sex differences in consistency of moral orientation were found. This outcome is

in line with Wark and Krebs (1997) who reported consistency on adjacent scores by only two

participants.

Content Analysis of impersonal dilemmas

All the impersonal dilemmas were classified into the categories: antisocial,

philosophical, and prosocial based on the topics which were raised by the participants

themselves. An initial analysis of the descriptive statistics revealed that 28 participants (17

females & 11 males) reported prosocial dilemmas (48.3%), 52 participants (28 females & 24

males) reported antisocial dilemmas (89.7%), and 36 participants (15 females & 21 males)


reported philosophical dilemmas (62.1%). Three univariate Anovas2 (combining both gender

& gender role factors with each type of impersonal dilemma) were carried out to determine

whether the content of the dilemmas influenced moral orientation scores. Although the three

groups differ on the number of males/females, gender role, and number of dilemmas, each

group had a representative number of each level of the factors, which enabled basic

requirements for statistical analysis. The analyses revealed no significant effects at all levels.

These results may have been different with a more strict statistical power control (e.g. an

equal number of dilemmas in each group). Nonetheless, the lack of significant findings also

suggest that moral orientation scores of the impersonal dilemmas represented a true variance

of care responses versus justice responses rather than the tendency to report specific types of

impersonal dilemmas.

Moral Reasoning – consideration of consequences

A next set of analyses turned to consider the proportion of judgements that were led by

a consideration of consequences. Figure 2 shows the proportion of judgements led by

consequences for both pro- and antisocial personal and impersonal dilemmas.

--Insert Figure 2 about here--

Two paired-sample t-tests were carried out to detect whether the difference between the

means is statistically significant. Results revealed a significant difference between the

impersonal-antisocial dilemmas and the personal-antisocial dilemmas: t (51) = 5.51, p<0.001,

but no significant difference between the impersonal-prosocial dilemmas and the personal-

prosocial dilemmas.

2 The p-value for the univariate Anovas was adjusted to 0.0167 in order to compare them to

the initial 3-way Anova since participants were instructed to report only two impersonal

dilemmas.


These results support current research by Krebs and his colleagues by providing an

insight to an important aspect of moral decision-making. The results suggest that the effect of

consequences on judgement when one is faced with when making a decision (a personal

dilemma) is stronger than the effect of consequences one considers other people to have when

they are faced with moral conflicts (an impersonal dilemma). However, this is true only with

regard to antisocial conflicts – it does not hold for prosocial conflicts.

One further analysis, examining the relationship between participants' consideration of

consequences and moral orientation was not significant.

DISCUSSION

The present study explored two aspects in relation to moral decision-making. The first

part (moral orientation) examined Gilligan’s theory of moral orientation and its relationship to

dilemma type, gender and gender role. The second part (moral reasoning) explored a new

aspect in recent research: consideration of consequences, focusing on whether consequences

to the self differ from consequences to other people in the influence they have on decision-

making.

Moral Orientation

The results of this study are not consistent with Gilligan’s hypothesis about moral

orientation. Gilligan and Attanucci (1988) acknowledged that individuals can employ both

justice and care orientations but they also claim that only one (either justice or care) prevails

across an individual’s thinking. This study however, failed to find a prevailing orientation

across all four dilemmas. Not only do participants include both care and justice judgements in

their dilemmas, but the majority (all but three) were inconsistent with the orientation they

presented to address each dilemma. This outcome supports Wark and Krebs’s (1997) study, in

which they claim to have found poor consistency of moral orientation across four personal

dilemmas.


Gilligan (1982) also proposed that the type of orientation is related to gender, assuming

that males tend to focus on the justice orientation whereas females tend to focus on the care

orientation. Like previous research (e.g. Walker et al.. 1987) this study fails to support this

claim. On average, females focused on care-based judgements more than males did but this

did not result in a statistically significant difference between the two gender groups. However,

the overall mean (57%) for orientation suggests that all participants (regardless of gender)

expressed slightly more care judgements than justice judgements. Furthermore, the

consistency measure did not yield gender differences, which suggests that males were as

inconsistent in their moral orientation as females were. In a similar vein gender did not

produce a significant main effect or significant interactions, which leaves one to conclude that

as far as this study was able to show moral orientation was not related to gender.

This study, nevertheless, produced significant results regarding type of dilemma and

gender role. A main effect for dilemma type qualified by an interaction between dilemma and

gender role was found. A post-hoc analysis revealed the exact location of these significant

differences: between the (a) prosocial and the antisocial dilemmas and (b) the prosocial and

impersonal dilemmas at the femininity level and at the androgyny level. Similar to previous

studies which included gender role as a factor (Sochting, Skoe, & Marcia, 1994) this study

supports the claim that gender role may serve as a better predictor of moral orientation than

gender alone. Moreover Ford and Lowery (1986) found the significant results to be at the

femininity level, which has been replicated by this study.

Unlike Ford and Lowery (1986), this finding was not related to gender and extends to

generalise gender role as an important factor by finding significant comparisons at the

androgyny level as well. Both androgyny and femininity gender roles were defined by a high

score on the feminine scale (Spence & Helmreich, 1978). The fact that dilemma type reached

significance at these levels only (only for the ‘androgynous’ and ‘feminine’ people) may

provide the answer for the distinction between the significant results of those scales and the

non-significant results of the masculine and the undifferentiated scales, which were both low

at femininity. In other words, one may need to have high level of femininity in order to show


significant differences between some types of moral dilemma (e.g. prosocial-antisocial &

prosocial-impersonal).

The overall pattern of results regarding type of dilemma is similar to earlier findings by

Wark and Krebs (1997). Wark and Krebs (1997) compared three types of dilemma (2

Kohlbergian, 2 real life-prosocial, & 2 real life- antisocial). They reported that the

Kohlbergian dilemmas pulled for justice orientation, the prosocial tended towards care

orientation, and the antisocial towards justice orientation. In the present study, results were in

a similar direction.

There is, however, a marked difference between the means of the Kohlbergian

dilemmas (Wark & Krebs, 1997) and the impersonal dilemmas (the present study). The notion

that Kohlberg’s (philosophical) dilemmas evoke justice-based judgements is not new. Indeed,

it was one of the main criticisms against Kohlberg’s model (Gilligan 1982). However, the use

of impersonal dilemmas in the past has led to some interesting outcomes. Wark and Krebs

(1996) reported that the philosophical impersonal dilemmas evoked a similar level of justice

to that evoked by Kohlberg’s dilemmas, and that they evoked a significantly lower level of

care than the antisocial impersonal dilemmas. These findings appear inconsistent with the

outcomes of the present study but this inconsistency may reflect different tendencies in the

scoring procedure. The means of the prosocial impersonal dilemmas and the antisocial

impersonal dilemmas are fairly similar to the means of the personal prosocial and antisocial

dilemmas, and share similar locations on the moral orientation scale (antisocial towards

justice and prosocial towards care). The overall mean of the impersonal dilemmas seems to

reflect the grouping of the various types of impersonal dilemmas (prosocial, antisocial, &

philosophical). However, the statistical analyses suggest that there was no particular

influence from any type of dilemma on the overall mean of the impersonal dilemmas. There

was no significant difference between the impersonal dilemmas on moral orientation scores.

Methodologically, the need to control types of moral dilemma (for both personal &

impersonal dilemmas) cannot be over emphasised. Wark and Krebs (1996) found that females

tend to report more prosocial dilemmas whereas males tend to report more antisocial


dilemmas. These tendencies affect moral orientation scores – the present study demonstrated

that prosocial dilemmas evoke more care judgements whereas antisocial dilemmas evoke

more justice judgements. These gender-related patterns of reporting have not been

demonstrated by this study, yet they may explain the outcome of Gilligan and Attanucci’s

(1988) study in which type of dilemma was not held constant. Moreover, Wark and Krebs

(2000) found that women report more prosocial real life dilemmas, perhaps because women

consider prosocial dilemmas to be more significant as these dilemmas elicit most guilt (Wark,

1998), whereas men experience more antisocial (transgression) types of conflict in real life.

Thus, Gilligan’s notion of moral orientation may be embedded in life experience rather than

to any particular gender group per se.

Moral Reasoning – consideration of consequences

The results concerning the second part of this study support to some extent some

theoretical assumptions based on research by Krebs and his colleagues (in particular Krebs et

al., 1997). The statistical analysis revealed a significant difference between consideration of

consequences to the self and consideration of consequences to others but only as far as

antisocial dilemmas were concerned. When type of dilemma was held constant for both the

personal and the impersonal dilemmas, the prosocial dilemmas did not yield a significant

difference. In fact both impersonal-prosocial and personal-prosocial dilemmas had identical

means, which suggests that people tend to consider the outcome of their decisions and actions

as important with regard to themselves and others equally when discussing moral dilemmas

concerning prosocial issues.

The difference in the results between the prosocial dilemmas and the antisocial

dilemmas implies that people tend to regard consequences to themselves as highly important

compared to consequences to others when discussing antisocial issues. By way of a contrast,

consequences of prosocial dilemmas are regarded as important whether they relate to the self

or to others. Perhaps this is not such a great surprise bearing in mind that antisocial issues are


closely related to law and punishment whereas prosocial issues are to a large extent related to

one’s willingness to help another.

Krebs et al. (1997) provide a constructive distinction between moral conflicts in terms

of their anticipated consequences: (a) approach – approach conflicts (“should I spend more

time with my boyfriend or my friends?”), (b) approach – avoidance conflicts (“should I lend

money to my friend or avoid taking responsibility for him?”), and (c) avoidance – avoidance

conflicts (“should I lie to my landlord or face eviction from my flat?”). The conflicts

discussed in our antisocial dilemmas involved at least one avoidance aspect (e.g. underage

drinking versus getting caught, facing condemnation versus feeling guilty, etc). It would be

valid to assume that decisions, which involve consequences to the self that one is trying to

avoid, will have more effect on one’s moral decision than decisions entailing consequences

that others may try to avoid.

This last point brings us back to potential dissonance between should and would that

was described by Krebs et al. (1997). In the same way that this distinction explains the

inconsistency between reasoning of hypothetical dilemmas and reasoning of real life

dilemmas, it may explain the inconsistency between justification of others’ reasoning (when

one is an observer) and the justification of one’s own reasoning. The outcomes of antisocial

conflicts (e.g. law breaking, being unfair or unjust) often contrast one’s own morality

standard. Denton and Krebs (1990) found that people tend to consider themselves to be more

moral than other people, a phenomenon they named ‘the self-righteous bias’. This

phenomenon ties in with the current findings. According to Krebs et al. (1997) people invest

in their moral identities, which in return affect their moral decisions. In situations where

people behave inconsistently with their moral identities (e.g. antisocial type situation) they are

faced with negative outcome (physical or mental) and negative reputation, which motivate

them to reduce the inconsistency between their belief about their own moral identity and how

they have been perceived by society (judicial system, family relatives, friends, etc). This

attempt to reduce negative reputation of one’s moral identity is manifested in moral dilemmas


in the form of justification of behaviour, which as research showed, involves low stage moral

structures (Denton & Krebs 1990; Krebs et al, 1991; Wark & Krebs 1997; Krebs et al. 1997).

This is a crucial observation for research on moral decision-making as it can only be

explored by real life personal dilemmas where participants are asked to justify their own

experiences. It follows then that in the present study participants may have felt the need to

justify their own behaviour and reduce the inconsistency of their own moral identity, which

led them to consider the consequences of their decisions/actions in a way that affected their

decisions. In other words, participants regretted acting in a certain way and therefore justified

their behaviour by considering the consequences of their actions in order to avoid similar

outcomes in the future. However, when asked to discuss others’ moral decisions in antisocial

situations the need to justify others’ behaviour in terms of its consequences was less

important3. On the contrary, prosocial behaviour educes a positive moral reputation that is

more consistent with people’s moral identity (the ‘self righteous’ bias), and may validate or

even improve one’s perception of oneself (Krebs et al. 1997).

Conclusions

These findings do not corroborate Gilligan’s theory of moral orientation. Participants

rarely held one orientation across all dilemmas. Moreover gender was related to neither

justice nor care orientations. Significant comparisons were found between the prosocial and

the impersonal dilemmas and between the prosocial and the antisocial dilemmas for both

feminine and androgynous gender role groups. This outcome suggests that although people, in

general, do not hold a particular moral orientation, ‘feminine’ and ‘androgynous’ people score

3 Although it must be remembered that our definition of “personal” dilemmas were those

which directly involved an individual and made that individual take some form of action.

Thus it may be true that when an individual makes a decision not to act prosocially (although

not necessarily antisocially) there may be some justification in terms of the potential

consequences for the self.


significantly higher percentages of care-based responses when discussing prosocial dilemmas

compared with impersonal or antisocial dilemmas. The latter focuses the attention on the

effect of external sources (e.g. type of dilemma) on moral decision-making. The

acknowledgement of such interaction between external and internal (e.g. gender role) sources

of variation on moral decision-making is crucial to the understanding of how people judge

real life moral conflicts.

Results also indicate, as might have been anticipated, that people consider the

consequences of their decisions when they discuss moral conflicts. Furthermore, the

consequences of moral decisions seem to have more influence when people discuss personal-

antisocial conflicts rather than impersonal-antisocial conflicts. This difference was not evident

between the prosocial-personal/impersonal dilemmas. It may be that when people discuss

personal-antisocial (e.g. violations of rules, laws, or fairness) dilemmas they seek to resist

adopting a negative reputation. The therefore justify their own behaviour with a higher

percentage of consideration of consequences in their attempt to view themselves more

positively (and enhance their moral identity). This pressure disappears when people discuss

impersonal-antisocial dilemmas because the need to justify other people’s behaviour in a

positive way is less strong and has less influence on judgements and reasoning. More so, it is

not apparent in prosocial dilemmas because this type of behaviour has a positive reputation

and entails a positive moral identity. Consequently, the outcomes of people’s prosocial

behaviour have less influence on their reasoning.

It is also possible to see some ways in which the current findings might inform work in

moral education. For instance, we see in the results of the current study a link between

internal (gender role) and external (dilemma type) factors in making moral judgements. There

is also now strong evidence to suggest that the underlying motivations for moral judgements

differ according to the type of dilemma under consideration. In view of this, educators need to

consider whether it might be appropriate to employ different strategies for encouraging

mature moral reasoning with respect to pro- and antisocial behaviour. Moreover, the influence


of internal factors such as gender role points to a need to gear any educational interventions to

the needs of specific individuals.

As the current study has demonstrated, another important factor in making moral

decisions is a consideration of the consequences of actions. Such consideration appears at its

most influential when reasoning about one’s own response to antisocial dilemmas (having

done or doing something wrong). As has already been indicated, the motivation to maintain a

reputation or positive moral identity not least, one might imagine, amongst one’s peers, could

explain findings here. However, it was not the case that participants in this study were

motivated to gain a positive (prosocial) self-identity but rather that they were motivated to

defend themselves against acquiring a negative (antisocial) one. This is an issue that is less to

do with moral education and more to do with the values we encourage as a society; behaving

“morally” is less about prosocial behaviour and more about not committing antisocial

behaviour. But a greater emphasis, in the schoolroom and beyond, on the social merits of

prosocial behaviour might just encourage more of it.

Finally, when considering antisocial dilemmas from an “abstract”, impersonal

perspective people imagine the consequences of an action as less important than when they

consider a similar event from a first personal perspective. There is, it would seem, a

separation of the actual from the theoretical here (at least in the reports of our participants).

Further studies from an educational perspective could help to identify whether encouraging a

child or adult to reflect on their own experiences and past, real-life moral decisions might

trigger forms of reasoning that are better suited to helping individuals make more mature

decisions in future, real-life moral dilemmas.

Altogether this study overcomes methodological problems in earlier studies and

confirms previous findings in terms of the effect of dilemma type and gender role on moral

orientation. It also points to the importance of individuals’ consideration of consequences of

their actions in judging real life moral dilemmas. Further investigations are needed to clarify

the role of this new aspect of moral reasoning that appears to be important in moral decision-

making. Such investigations could help in the development of interactional models of moral


reasoning that account for the interplay between internal (e.g. gender role) and external (e.g.

dilemma type) influences on everyday moral reasoning.


REFERENCES

Bussey, K. & Maughan, B. (1982). Gender differences in moral reasoning. Journal of

Personality and Social Psychology, Vol. 42, No. 4, pp.701-706.

Carpendale, J.I. & Krebs, D.L. (1995) Variations in Level of Moral Judgment as a Function of

Type of Dilemma and Moral Choice, Journal of Personality, Vol 63, No. 2, pp. 289-

313.

Denton, K. L. & Krebs, D. L. (1990) From the Scene to the Crime: The Effect of Alcohol and

Social Context on Moral Judgment, Journal of Personality and Social Psychology, Vol.

59, No. 2, pp. 242-248.

Ford, M.R. & Lowery, C.R. (1986) Gender Differences in Moral Reasoning: A Comparison

of the Use of Justice and Care Orientations, Journal of Personality and Social

Psychology, Vol 50, No. 4, pp.777-783.

Gilligan, C. & Attanucci, J. (1988) Two Moral Orientations: Gender Differences and

Similarities, Merrill – Palmer Quarterly, Vol. 34, No. 3, pp. 223-237.

Gilligan, C. (1982) In a Different Voice: Psychological Theory and Women’s Development.

Cambridge, MA: Harvard University Press.

Kohlberg, L. (1969) ‘Stage and Sequence: The Cognitive-Developmental Approach to

Socialization’, In D. Goslin (Ed.), The Handbook of Socialization Theory and

Research (pp.90-125). Chicago: Rand McNally.

Kohlberg, L. (1984). Essays on Moral Development, Volume 2: The Psychology Of Moral

Development. New York: Harper & Row.

Krebs, D.L., Denton, K.L., Vermeulen, S.C., Carpendale, J.I., & Bush, A. (1991) Structural

flexibility of Moral Judgment, Journal of Personality and Social Psychology, Vol 61,

No. 6, pp.1012-1023.

Krebs, D.L., Denton, K.L., & Wark, G. (1997) The Forms and Functions of Real-life Moral

Decision-making, Journal of Moral Education, Vol. 26, No. 2, pp.131-145.


Krebs, D.L., Vermeulen, S.C, Denton, K.L., & Carpendale, J.I. (1994) Gender and

Perspective Differences in Moral Judgment and Moral Orientation, Journal of Moral

Education, Vol 23, No. 1, pp.17-26.

Leman, P.J. (2001) ‘The Development of Moral Reasoning’, in C. Fraser, B. Burchell, D. Hay

& G. Duveen (Eds.), Introducing Social Psychology, Cambridge: Polity Press.

Lyons, N.P. (1983) Two Perspectives: On Self, Relationships, and Morality, Harvard

Educational Review, Vol. 53, No. 2, pp. 125-145.

Milgram, S. (1963). Behavioural Study of Obedience. Journal of Abnormal and Social

Psychology, Vol. 67, pp.57-76.

Pratt, M.W., Golding, G., Hunter, W., & Sampson, R. (1988) Sex Differences in Adult Moral

Orientation, Journal of Personality, Vol. 56, No. 2, pp.373-391.

Sochting, I., Skoe, E. E. & Marcia, J. E. (1994). Care-Oriented Moral Reasoning And

Prosocial Behaviour – A Question Of Sex-Role Orientation. Sex Roles, Vol. 31, Nos.

3-4, 131-147.

Spence, S.T. & Helmreich, R.L. (1978) Masculinity and Femininity – Their Psychological

Dimensions, Correlates, & Antecedents. Austin: University of Texas Press.

Walker, L.J. (1984) Sex Differences in the Development of Moral Reasoning: A Critical

Review, Child Development, Vol. 55, pp.677-691.

Walker, L.J., de-Vries, B., & Trevethan, D. (1987) Moral Stages and Moral Orientation in

Real-Life and Hypothetical Dilemmas, Child Development, Vol. 58, pp. 842-858.

Wark, G.R. & Krebs, D.L. (1996) Gender and Dilemma Differences in Real-Life Moral

Judgment, Developmental Psychology, Vol. 32, No. 2, pp.220-230.

Wark, G.R. & Krebs, D.L. (1997) Sources of Variation in Moral Judgment: Toward a Model

of Real-Life Morality, Journal of Adult Development, Vol. 4, No. 3, pp.163-177.

Wark, G.R. & Krebs, D.L. (2000) The Construction of Moral Dilemmas in Everyday life,

Journal of Moral Education, Vol. 29, No. 1, pp. 5-21.

Appendix 1: The Probe Questions Form 1) What did you see to be the issues involved at the time? 2) What made it a moral conflict? 3) What options did you or the person/people involved in the dilemma consider? Why? 4) Do you think you (or the person/people involved) did the right thing? 5) Why / Why not? 6) How about the other people or person involved? 7) What was each person’s responsibility? Why? 8) In general, should people help one another? why? (please relate your answer to each dilemma) 9) Should people always follow their moral principles? (please relate to each dilemma) 10) Are there any circumstances where it is right not to do the ‘right’ thing? (how is that relate to each dilemma) 11) Following your (personal) or other’s (impersonal) experience, did you change your point of view on the matter? 12) In what way and why? 13) In cases where some sort of relationship is described in your dilemma: What is the most important thing that each party should be concerned with and why? For impersonal dilemmas only: 14) How would you have acted if you found yourself personally involved in this situation and why? Please describe all the perspectives that you think are involved. 15) Using the scale from the first questionnaire, please rate the importance of the conflict in your life at that time: Not at all important A….B….C….D….E Very important For personal dilemmas only: 14) Under what circumstances would you have acted differently and why? 15) If in the future you will come across a similar situation, how do you think you will act and why? 16) What would you have done / How would you have acted if you were the other person or people involved and why? 17) Do you think your judgement of the situation would have been different if you were not personally involved? 18) In what way and why? Using the scale from the first questionnaire, please rate: 19) The degree of difficulty you have experienced in making the decision about what to do: Not at all difficult A….B….C….D….E Very difficult 20) How constrained you felt when you made your decision (and acted accordingly): Not at all constrained A….B….C….D….E Very constrained

Table 1. Mean proportions of judgements that were predominantly care-based

Gender role Dilemma

type Masculine Feminine Andro-

gynous Undiff-

erentiated Total

Impersonal Gender Male .56

.35

.32

.67

.50

Female .62

.55

.45

.61

.55

Antisocial Male .44

.45

.50

.35

.44

Female .40

.58

.41

.44

.47

Prosocial Male .60

.80

.80

.75

.70

Female .80

.83

.84

.66

.78

Figure 1. Total means for moral orientation score of all levels of all factors (expressed as percentages)

Moral orientation score

0 20 40 60 80

Dilemma - impersonal

antisocial

prosocial

Gender - men

women

Gender role - masculine

feminine

androgynous

undifferentiated

Figure 2. Means proportion of judgements that were led by consideration of consequences for

impersonal and personal, antisocial and prosocial dilemmas

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Prosocial Antisocial

Type of dilemma

Impersonal

Personal

Moral decision-making in real life: factors affecting ... - CORE

Documents